prgs-docs-spreadsheetkvc VS pmemstat

There are some esoteric curses based spreadsheets which likely solve the want-it-in-a-terminal problem w/o any coding (e.g., hanishkvc/prgs-docs-spreadsheetkvc: A simple commandline spreadsheet based on curses and python, with support for authenticated encryption based csv files.).

1.3GB is being used w/o running much, not even a desktop apparently, and btop is not helping identify the RAM eaters. BTW, I prefer pmemstat: Proportional Memory Status - shows memory consumption of Linux process by PID and groupings for 1st assessment of memory issues if you'd care to show that. That does seem high and it would be interesting to know why (although probably just academic).

There is no right answer to "what should ram usage be?"; however, for most desktop/laptop users, browsers and their tabs are the primary consumer of app memory (often starting at a few GB); DEs typically use 700MB, more or less. A tool that I prefer to use to see more exactly how much memory apps are using is pmemstat.

You are probably not going to see much difference in distros ... the browser is typically your pig and the differences in distros is in the noise. 4GB is rather tight. BTW, I prefer a "proportional memory" tool to analyze memory use of apps (e.g., pmemstat), but as I said, for most people, their browser is almost always the memory pig.

I use the Fedora KDE spin (with Xorg) and below are my numbers from pmemstat (which I use for accuracy/grouping of app memory even though it can be a bit piggy itself); I have vscode, a media server, two browsers, etc, running and the system is loafing. I actually disable snaps and flatpaks (although that may not matter). To be sure, Fedora runs comparably to other distros that I've used. Anyhow, it is hard to spot problems w/o details.

I'm not sure if you can find an official 3rd party python wrapper package to ease using curses for your exact need; but here is an roll-your-own example: the Window.py in pmemstat. You can see (from its README illustrations) that app has a fixed header, scrollable rows below, and commands to select rows, re-order the data, etc.; that is, I think it vaguely resembles what you want (and, if nothing else, the amount of code might suggest the effort of going that route).

Now, having said that, most tools don't even try to report accurate memory use by apps because it is tough. For example, this tool shows proportional memory used by the apps (disregarding cache, etc): pmemstat. Apps used both shared and unshared memory; fairly counting memory which is shared between apps divides it between all the apps sharing it proportionally. Anyhow, if you actually have a problem (e.g., apps are consuming too much memory and you want to know why), then a proportional memory tool helps greatly.

pmemstat: Proportional Memory Status - shows memory consumption of Linux process by PID and groupings. It mainly answers the question, "where is my memory going?", and the answer is usually my web browser(s) ;-). An example:

That is a very generic statement. There are lots of ways to start the fault isolation to a process. You might try pmemstat: Proportional Memory Status to narrow down the culprit, category of memory, etc) to the process at least. Of course, if there is a kernel memory leak, no process will be the problem.

Aside: my personal contribution to understanding memory use is pmemstat - Proportional Memory Status. Here is an example:

IMHO, the most meaningful per process memory numbers are provided by smem (see How to install and use the smem memory reporting tool in Linux). But I prefer the non-standard pmemstat because of its roll-ups and looping features, etc.); pmemstat and smem agree to a reasonable extent.